<?xml version="1.0" encoding="ISO-8859-1"?>
<metadatalist>
	<metadata ReferenceType="Conference Proceedings">
		<repository>sid.inpe.br/sibgrapi@80/2007/10.02.15.36</repository>
		<metadatarepository>sid.inpe.br/sibgrapi@80/2007/10.02.15.36.16</metadatarepository>
		<site>sibgrapi.sid.inpe.br 802</site>
		<citationkey>DonattiWürt:2007:MeOrIn</citationkey>
		<author>Donatti, Guillermo S.,</author>
		<author>Würtz, Rolf P.,</author>
		<affiliation>Institut für Neuroinformatik, International Graduate School of Neuroscience, Ruhr-Universität Bochum</affiliation>
		<affiliation>Institut für Neuroinformatik, International Graduate School of Neuroscience, Ruhr-Universität Bochum</affiliation>
		<title>Memory Organization for Invariant Object Recognition and Categorization</title>
		<conferencename>Brazilian Symposium on Computer Graphics and Image Processing, 20 (SIBGRAPI)</conferencename>
		<year>2007</year>
		<editor>Gonçalves, Luiz,</editor>
		<editor>Wu, Shin Ting,</editor>
		<booktitle>Proceedings</booktitle>
		<date>Oct. 7-10, 2007</date>
		<publisheraddress>Porto Alegre</publisheraddress>
		<publisher>Sociedade Brasileira de Computação</publisher>
		<conferencelocation>Belo Horizonte</conferencelocation>
		<keywords>Computer Vision, Theoretical Neuroscience, Neuroscience.</keywords>
		<abstract>The integration of bottom-up with top-down object processing has always been a topic of major concern in computer vision. However, while a lot is known about feature extraction, the knowledge-driven aspect of perception has been recognized as important, but hard to probe experimentally and difficult to implement in computer vision systems. How object knowledge must be organized so that it supports scene perception and can be acquired automatically is a research problem of outstanding significance for the biological, the psychological, and the computational approach to understand perception. The present work aims to develop an object memory model which can provide fast retrieval and robust recognition and categorization. The underlying data structure is inspired by the neural network structure of the human brain, connecting similar object views with excitatory synapses and using inhibitory synapses to separate different ones. The insights derived from building such a computational theory and the properties of the resulting model have implications for strategies and experimental paradigms to analyze human object memory as well as technical applications for robotics and computer vision.</abstract>
		<language>en</language>
		<tertiarytype>Technical Poster</tertiarytype>
		<format>On-line</format>
		<size>59 Kbytes</size>
		<numberoffiles>1</numberoffiles>
		<targetfile>sibgrapi_donatti_final.pdf</targetfile>
		<lastupdate>2007:10.02.15.45.49 sid.inpe.br/banon/2001/03.30.15.38 administrator</lastupdate>
		<metadatalastupdate>2020:02.19.03.06.19 sid.inpe.br/banon/2001/03.30.15.38 administrator {D 2007}</metadatalastupdate>
		<mirrorrepository>sid.inpe.br/sibgrapi@80/2007/08.02.16.22</mirrorrepository>
		<usergroup>sdonatti@neuroinformatik.rub.de administrator</usergroup>
		<visibility>shown</visibility>
		<transferableflag>1</transferableflag>
		<hostcollection>sid.inpe.br/banon/2001/03.30.15.38</hostcollection>
		<contenttype>External Contribution</contenttype>
		<lasthostcollection>sid.inpe.br/banon/2001/03.30.15.38</lasthostcollection>
		<url>http://sibgrapi.sid.inpe.br/rep-/sid.inpe.br/sibgrapi@80/2007/10.02.15.36</url>
	</metadata>
</metadatalist>